It all comes down to sample size.
If you win over a small sample size, the results are insignificant. You might be a winning player, you might be a losing player or you might be a break-even player.
But if you win over a large sample size, then the results are much more significant and you can honestly consider yourself a winning player.
I'll give you an example using my own win rates:
As you can see, I am winning at 2NL ($0.01/$0.02 blinds) but losing at 5NL ($0.02/$0.05 blinds). But does that make me a winning player at 2NL and a losing player at 5NL? Look at the sample size. I cannot say whether I am a winning, losing or break-even player at 5NL yet, since I've played under 2k hands and have only lost 640bb, whereas at 2NL, I've played 48k hands and have won 2700bb, so my 2NL win rate is much more significant/accurate than my 5NL win rate.
So my advice is to keep track of all your wins and losses and keep track of the sample size (a HUD will do this automatically). Once your sample size reaches 10k+ hands, you'll be able to (roughly) assess whether you're a winning player or not. Until then, happy grinding